Effect of pregnancy intention, postnatal depressive symptoms and social support on early childhood stunting: findings from India

Childhood stunting, defined as height for age z-score below -2sd of the median of reference population by the World Health Organization, is remains one of the important public health concern throughout the developing countries [1]. According to United Nation Children’s Fund (UNICEF), it has been estimated that worldwide about 165 million children are stunted in 2012 and India alone accounts for 38 % of global burden of stunting in the world [2]. In terms of incidence, India accounts for more than 60 million stunted children worldwide [2]. Recent data from third Indian National Family Health Survey (INFHS) suggest that about 45 % children under age 3 year are stunted [3]. Such a high prevalence of stunting is a matter of serious concern because stunted children are associated with weaker immune system, higher risk of developing diarrheal disease, acute respiratory infection, delay in motor skills, cognitive, and social development during childhood [4‐12] and more likely to suffer from high blood pressure, obesity, diabetes and heart disease during adulthood [6, 13‐15]. Another concern related to stunting is it passes from one generation to another as a grim inheritance [16, 17].

Studies have identified several risk factors such as birth size, mother’s education, mother’s age at birth of the child, consumption of iron folic acid tablet during pregnancy and socio economic condition for childhood stunting in developing countries [18‐20]. Studies from India also reported that birth size, mother’s education, antenatal check-up during pregnancy, serious illness, drinking water, toilet facility, economic status, place of residence are statistically associated with childhood stunting [21‐23]. A study by World Health Organization (WHO) suggests that about 50 % of childhood stunting in India is related with repeated diarrhoea or intestinal worm infection from unsafe water and poor sanitation or hygiene [24]. However, very few studies have analysed the cumulative effect of both pregnancy intention and postnatal depressive symptoms on early childhood stunting simultaneously in developing countries and perhaps there was no such study in India.

Pregnancy intention may influence child health outcomes through increased level of depression, change in behaviour in parenting, had a greater risk of smoking, consumption of alcohol, and were less likely to attend antenatal care and to take iron folic tablets during pregnancy [25‐28]. Which in turn increase the risk of preterm birth, low birth weight, and of not receiving sufficient resources for healthy development [29, 30]. Existing literature on the association between pregnancy intention and childhood stunting has been inconsistent. A number of studies from developing countries reported that pregnancy intention is linked to childhood stunting [22, 31‐33]. On the other hand, a few studies reported that pregnancy intention is not associated with childhood stunting in developing countries [31, 33]. A study using cross-sectional data from INFHS, reported that unintended birth was disadvantage in terms of childhood stunting [22]. However, the majority of these studies did not adjust for psychosocial factors such as mental depression and social support.

Another subject worth exploring childhood stunting is postnatal depressive symptoms. In the context of developing countries poor maternal mental health, in particular, maternal depression may be a significant risk factor for poor growth during early childhood. In addition to depression, women are particularly prone in postpartum period because of hormonal changes associated with childbirth and stressor related to parenting [34‐36]. The combination of women vulnerability to mental depression and responsibility for child care in developing countries means that maternal mental health could have a significant effect on growth during early childhood [37]. A number of previous studies from developing countries have shown that maternal depressive symptoms were significantly associated with childhood stunting [35, 38‐43]. A study from Goa (India) also shows that children of the depressed mother were 2.3 times more likely to be stunted than children of non-depressed mothers [41]. A similar finding was also reported by another study from the rural area of Tamilnadu (India) [44]. However, these studies in India were based on small sample sizes and lacked representativeness.

The association between pregnancy intention, maternal depression, social support and child health outcome is undoubtedly complex [45‐48]. Unintended pregnancy may increase a women’s exposure to the psychosocial stressor, decrease social support provided to her by the partner, increase her level of depressive symptoms and decrease her overall life satisfaction [46]. However, social support has potential to play protective effect by buffering the impact of life stress on the well-being of the women during pregnancy [49‐52]. Studies have shown that social supports play a buffering role from stressful life events by providing resources, support and strength during pregnancy [46, 53‐57]. In an analysis of YLS data from four developing countries, De Silva and Harpham [58] found that support from the individual was linked to child nutritional status in Ethiopia, Peru and Vietnam. The only one study from Ethiopia reported that effect of maternal depression on child birth outcome was mediated by social support [57]. However, no such study exist from India. Therefore, in this study, we examined whether women who reported unintended pregnancy or who experienced postnatal depression are at higher risk of having stunted children in India. We hypothesized that effect of pregnancy intention and maternal postnatal depressive symptoms on childhood stunting was mediated by social support.

The characteristics of the study population are shown in Table 1. The result shows that about 8 % of women reported that their pregnancy was unintended or mistimed while the prevalence of postnatal depressive symptoms was about 30 %. The prevalence of stunting was about 27 % among children aged 5 to 21 months. A higher proportion of children were stunted of whose mothers had unintended pregnancy (40 %) than those who had intended pregnancy (26 %). Similarly, childhood stunting was also higher among those whose mothers had symptoms of postnatal depression (35 %) than those who had no symptoms of postnatal depression (24 %). The percentage of stunted children did not vary substantially according to the social support.

Prevalence of stunting was higher among children of small birth size than average/above average birth size (38 % against 23 %), those who born preterm compared with term babies (45 % against 26 %) and those who suffered from serious illness/injury during early childhood compared to those who did not suffer from any illness/injury (34 % against 25 %). A higher proportion of children were stunted of those women who had less schooling and who had the lower age at birth of the child. Likewise, the prevalence of stunting was higher in children, among mothers who did not go for the antenatal check-up, consume less than 90 IFA tablet and had less than 2 TT injections during their pregnancy. The prevalence of stunting also varied substantially according to the socioeconomic status of the household and rural-urban residence. Prevalence of stunting was higher in children from household using non-improved toilet facility, belongs to poor wealth quintiles and those who are living in rural area. Results of the logistic regression model to examine the effects of pregnancy intention, postnatal depressive symptoms and social support on childhood stunting are shown in Table 2. Results of unadjusted logistic regression model shows that among socio-demographic and psychosocial factors; pregnancy intention, postnatal depressive symptoms, birth size, preterm birth, serious illness, mother’s education, mother’s age at birth of child, education of household’s head, wealth index, toilet facility and place of residence were statistically associated with childhood stunting. The result of adjusted logistic regression model also shows that pregnancy intention and postnatal depressive symptoms were statistically associated with childhood stunting, after controlling social supports and well-known determinants of stunting. Children born after unintended pregnancy were 1.76 times (AOR- 1.76; 95 % CI: 1.25–2.48) more likely to be stunted than children born after intended pregnancy. Likewise, children of those women who showed symptoms of postnatal depression were 1.53times (AOR-1.53; 95 % CI: 1.21–1.92) more likely to be stunted than non-depressed women. Although, in the adjusted model, pregnancy intention and postnatal depressive symptoms was statistically associated with childhood stunting while the magnitude of the effects of pregnancy intention and postnatal depressive symptoms on childhood stunting was attenuated. In order to examine whether the effect of pregnancy intention and postnatal maternal depressive symptoms on early childhood stunting were mediated by the presence of social supports, we used a series of multivariate logistic regression models. Multivariate models I and II shows that unintended pregnancy was significantly associated with childhood stunting even after controlling social support and other socioeconomic and residence related variables (Table 3). Likewise, multivariate models III and IV shows that postnatal depressive symptoms are statistically associated with childhood stunting after adjustment of social support and other socioeconomic and residence related variables (Table 4). Notably, the magnitude of the effect of pregnancy intention and postnatal maternal depressive symptoms on childhood stunting was not changed even after including the social support. Social support was not statistically associated with childhood stunting. Apart from these variables some other variables are also significantly associated with childhood stunting in the multivariate model. Children of small birth size than average or above birth size, preterm birth compare to the term baby, lower age of mother at birth of the child and belong to lower wealth quintiles were more likely to be stunted than their counterpart children (Table 2).

Our findings clearly suggest that mother’s pregnancy intention and postnatal depressive symptoms are significantly associated with early childhood stunting, after adjustment of the well-known determinants of child nutritional status in India. Finding indicates that unintended births were significantly more likely to be stunted than intended during early childhood. Several previous studies reported the inconsistent association between pregnancy intention and childhood stunting [31, 69]. However, findings of this study are consistent with previous studies from India [22], Bolivia [34], Peru [31] and Bangladesh [70]. Further, the study provides a significant association between maternal postnatal depressive symptoms and childhood stunting. Our findings are consistent with previous studies conducted in developing countries [40, 41, 43]. In contrast, to these studies a few studies have reported no association between depressive symptoms and childhood stunting in developing countries [71‐73]. However, the majority of these studies in developing countries did not adjust for psychosocial factors including pregnancy intention and social support. Moreover, findings indicate that effects of pregnancy intention and postnatal depression on early childhood stunting were not mediated by the social support. One of the possible reason for the lack of association between social support and childhood stunting may be reverse causality whereby women who are depressed are more likely to receive support from other individuals within their community, relatives and friends. Notably, Kawachi and Berkman [74] reported that protective effect of social support may not be uniform across society. They argue that social connection may increase the level of depression among women with low resources. A similar finding was also reported by Mitchella and LaGlory [75] in an impoverished community in the USA. Such a scenario is plausible among the poor biased YL sample. The study found significant interactions between pregnancy intention and maternal postnatal depression in predicting early childhood stunting. To the best of our knowledge, this is the first study of its kind that examined effect of pregnancy intention and maternal postnatal depression on childhood stunting in India.

This study has some limitations. First, only 8 % of the births were reported as unintended by the mothers in YLS. This figure is much lower than that obtained for Andhra Pradesh (India) from the INFHS 2005–06 (18.4 %). The difference in the estimates between the INFHS 2005–06 and YLS could be due to two reasons. First, the wordings of the questions in the two surveys are completely different. Notably, the questions on pregnancy intention in INFHS are far more comprehensive than the questions in YLS. Secondly, INFHS provides estimates for Andhra Pradesh whereas the YLS is restricted to few sentinel sites in Andhra Pradesh. Given the difference in the two estimates, the estimates of unwanted births in YLS can be safely taken as a lower bound of the actual estimate. Hence, the effects of pregnancy intention that we have obtained in our study would have been more compelling and YLS accurately captured the unintended births. Second, our study could not distinguish the separate effect of unwanted birth and mistimed birth on early childhood stunting due to data limitation. Third, our estimate of the effect of pregnancy intention on childhood stunting is likely to be lower than expected due to mortality selection. There is a higher chance of death during infancy among unwanted children as compared to wanted children [22, 23]. Our estimates of the effect of pregnancy intention on childhood stunting are also likely to be affected by differential reporting of pregnancy intention by the mother. Differential could be happened as data based on the retrospective report of intention and this method is subject to bias in recall and reporting. For example, during the survey mother may report that index child is intended at the time of conception, although at the time of conception child was unintended. Most researchers agreed that mother’s feeling and attitude regarding pregnancy intention change from unintended to intended following the births of child and hence reduces the effect of pregnancy intention on childhood stunting [32]. However, pregnancy intention was measured soon after birth (from 5 to 21 months), therefore, bias in recall and reporting should be minimal. Another concern related to quality and timing of the social support, we are only able to control for type of social support. However, quality and timing of the social support is also essential [76, 77].

Despite these limitations, our study has some strength. First, a large cohort of children included in the analysis and represents children of a wide range of family background. Second YLS is the only large scale available data that provides information on pregnancy intention, maternal postnatal depressive symptoms, social support and childhood stunting in India. Third, YLS uses a child focus mixed methods sampling approach, allow the examination of the complex interrelationship between pregnancy intention, postnatal depressive symptoms, social support and childhood stunting in India. Fourth, study includes both rural and urban areas, representing a range of regions, policy context and living condition that reflects the ethnic, geographical and religious diversity of the country. Finally, our study came out with some findings that may either lead to the formulation of new policies or may lead to the strengthening of the existing policies and programmes.

Our study provides strong evidence regarding the effect of pregnancy intention and postnatal depressive symptoms on early childhood stunting in India. The findings of our study have important policy implications. The prevalence of early childhood stunting is still serious health issue in India. Recent estimates from INFHS suggest that about 48 % of children under age five are stunted in India, which results in more than 60 million of stunted children. The risk of early childhood stunting is higher among women of unintended pregnancy and those with postnatal maternal depressive symptoms in India. Therefore, there is urgent need to identify the women with unintended pregnancy during antenatal care visit and incorporate the promotion of mental health into their national reproductive and child health programme.

Our study is based on a secondary dataset with no identifiable information on the survey participants. This dataset is available in public domain for research use and hence no approval was required from any institutional review board. The data can be downloaded from the website of the United Kingdom Data Archives University of Essex after creating account (https://​www.​ukdataservice.​ac.​uk/​). The data for the current study was downloaded from the afore-mentioned website after taking permission (I.D. No. 90978).

Not applicable.

The data which support our findings is contained within the manuscript.